A Process for Data Requirements Analysis. David Loshin Knowledge Integrity, Inc. October 30, 2007

Transcription

1 A Process for Data Requirements Analysis David Loshin Knowledge Integrity, Inc. October 30,

2 Agenda The criticality of business information Data requirements analysis Conducting stakeholder interviews Key data discovery artifacts Version 0.5 8/7/2007 2

3 Data Warehouse Information Perspective Business Decision and/or Action Knowledge Information Data Version 0.5 8/7/2007 3

4 Business Information Definition of Business Information Data that has been collected and organized to support management of business operations and decision making. Non transactional data e.g. counts Must be actionable Aggregation Users Delivery Detailed Operational Data Aggregated Management Data Summarized Internal & External Data Structured analytic data Staff and line managers Mid-level and senior managers Executive staff Special purpose marketing, business process analysis Queries and detail listings Summary reports and scorecards Dashboards Data mining, OLAP, Analytics, etc. Version 0.5 8/7/2007 4

5 Applications Driven by Business Information Vendor Analysis Customer Analysis Customer Profiling Targeted Marketing Personalization Customer Satisfaction Customer Lifetime Value Customer Loyalty Supplier Management Vendor Management Shipping Inventory Control Distribution Analysis Staff Productivity Call Center Utilization Call Center Optimization Production Effectiveness Business Productivity Defect Analysis Capacity Planning Asset Management Resource Planning Financial Reporting Risk Management Enterprise Reporting Business Information Behavior Analysis Trend Analysis Web Activity Fraud/Abuse Detection Participant Attrition Interaction Analysis Version 0.5 8/7/2007 5

6 Types of Business Information Key Performance Indicators (KPI) metrics for monitoring progress toward objectives Operational Performance Indicators (OPI) metrics for monitoring business processes, operations and workflow Leading Indicators (LI) metrics that contribute to rolledup KPIs or OPIs Supporting Metrics (SM) metrics supporting drill-through of KPIs, OPIs, and LIs Business Intelligence (BI) analytics to understand customer behavior, patterns, trends and segmentation Version 0.5 8/7/2007 6

7 Business Intelligence and Metrics Measures for setting objectives for strategic initiatives Measures for tactical activities Key Performance Indicators Operational Performance Indicators Leading Indicators Supporting Metrics Supporting Data

8 What is Data Requirements Analysis? The Data Requirements Analysis Process is a standard set of procedures for identifying the data needs of a Data Warehouse system. The steps are analogous to traditional requirements analysis, but focused on data rather than functional needs. Version 0.5 8/7/2007 8

9 Goals of Data Requirements Analysis

10 Goals of Data Requirements Analysis Strategic Requirements Information Requirements Technical Requirements Requirements Prioritization Fact/Qualifier Matrix Source Systems Source/Target Mapping Conceptual Data Model Version 0.5 8/7/

11 Data Warehouse Information Perspective Iterative, top-down approach Business decisions Data needed to support those decisions Interviews to identify information needs and metrics: Performance Management KPI Operations Management OPI Business Analysis BI Customer Performance CPI Analysis Specifications Data Discovery Data Assessment Testing & Validation Data Quality Control Version 0.5 8/7/

12 12

13 Driving Out Business Information Requirements Relevance Does the data: Support business processes? Provide metrics for OPIs and KPIs? Answer business questions? Enable correct measures? Reflect real world activities or state? Provide useful capture points, timeliness Value-Added Can the data improve: Operational visibility (e.g., document workflow and bottlenecks)? Efficiencies (e.g., in volume or processing times)? Decision making (e.g., expose patterns, trends, problems and opportunities)? Available Does data support: Captured information requirements? Timeliness for updates to support information requirements? Organization structure to support information navigation needs? Data sources Granularity Dimensionality

14 Data Discovery Data Discovery tasks focus on: Capturing business information requirements, business processes, and terminology Identifying and defining the source data sets Defining the data quality dimensions for these data elements. Deliverables Context Diagram Business Questions Fact/Qualifier Matrix Candidate Source Systems Source to Target Mapping DB Scenarios including Transformation Rules & Business Logic Version 0.5 8/7/

15 Version 0.5 8/7/

16 The Requirements Analysis Process Identify the business context Conduct stakeholder interviews Synthesize requirements Source to Target mapping

17 Identifying the Business Context - Activities 1. Identify relevant stakeholders 2. Review and summarize the overall goals, objectives for the data warehouse Acquire all project documentation Acquire all system documentation 3. Review and summarize the scope of the reporting platform s capabilities 4. Review and summarize impacts and constraints Version 0.5 8/7/

18 Identifying Business Context - Diagram BI Repository Metadata Business Rules ETL ODS Detailed level data Reference tables Querying Reporting Analytics Integration tables Users review business information via query/reporting/analytics tools

19 Conduct Stakeholder Interviews - Activities 1. Identify interview candidates 2. Prepare and schedule interviews 3. Conduct interviews 4. Summarize interview notes 5. Identify information gaps for follow-up Version 0.5 8/7/

20 Summarize Interview Notes Interviewee: Date/Time/Location: Question: Answer/Notes: Question: Answer/Notes: Version 0.5 8/7/

21 Synthesize Requirements - Activities 1. Create a reference workflow model 2. Develop and validate the Fact/Qualifier Matrix 3. Identify candidate source systems 4. Identify and Standardize Common Business Terms Version 0.5 8/7/

22 Business Process Modeling A business process is a coordinated set of activities intended to achieve a desired goal or produce a desired output product Models are designed to capture both the high level and detail of the business process controls inputs activity output actors

23 Shared Data Objects Interactions between activities depend on shared data: activity activity activity Instance values of common data types representing business facts communicate input and control during the business process Use the model to identify measurement points to be subjected to reporting and/or analysis

24 Work Flows and Tasks - Example Work Flow and Task Work Flow- A Event Trigger Event Status Measures Initiated Rejected Rework Complete Received Pending Cleared Complete Volume Start Date End Date Duration Task 1 Task 2 Work Flow- B Task 1 Task 2 Version 0.5 8/7/

25 Facts and Qualifiers Facts: specific business questions or discrete items of information to be tracked/monitored/reported Examples: counts, volumes, high water marks Qualifiers: conditions or dimensions used to filter or organize facts Examples: regions, owners, time Facts 800# calls processed Submissions received by mail Submissions returned Qualifiers Time Region Media Office X X X X X X X X s processed X X X

26 Common Business Terms - Glossary Term Address Verification Business Intelligence (BI) Repository CMM Count (Workload) Definition An employee verifies that the submitters address on file is correct for contact purposes. The BI Repository captures and maintains metadata (data about data). The metadata housed in the BI Repository consists of business and technical data names and definitions; information about where the described data resides in enterprise databases; and what database applications and reports use the data. The BI Repository is accessible through a web front-end application. Capability Maturity Model describes the principles and practices, through which organizations define, implement, measure, control, and improve their software processes and is intended to assist organizations in software quality and process improvement. For each Unit of Work, the counts associated at the task level for each data mart status (e.g., Receipt, Pending, Clearance, and Completion). Version 0.5 8/7/

27 Proposing the Target Model Evaluate the catalog of identified data elements Seek out the frequently created, referenced, modified, retired Assess object organizational structure Evaluate conceptual structures as they map to business process use Example: locations are composed of street, city, state, ZIP code Identify and resolve anomalies across data element sizes, types, formats Propose an object model Validate the object model within the information framework Validate the object model within the application framework

28 Source to Target Mapping - Activities 1. Identify and document source to target data mapping for each target element 2. Define the relevant transformation rules based on business scenarios 3. Identify the Candidate Reference Tables Version 0.5 8/7/

29 Source to Target Mapping Example Target Data Element Source Systems Source Table Source Data Elements ProcessType SYSA SUBTABA Process_Type Transformations Order_Count SYSA SUBTABB UID COUNT(SYSA.UID) LNAME WHERE SYSA.LNAME = SYSB W3 LASTNAME SYSB.LASTNAM E Version 0.5 8/7/

30 Data Assessment Tasks focus on: Identifying the best sources of data Assessing the quality of the data sources Identifying gaps in requirements versus availability Deliverables: Data assessment results Updated Source to Target Map Gap report

31 Data Assessment Version 0.5 8/7/

32 Data Profiling for Data Requirements Analysis Characteristics of data element metadata is critical for analysis Combination of artifact review and empirical analysis Data profiling can provide ground-truth evidence of consistency with metadata Provides insight into suitability of candidate sources to satisfy the target needs

33 Gap Analysis Template Source Data Target Data - Data Element Name System Table Column Data Type NULL Option % Null or missing Min Value Max Value Comments Requirements ID Version 0.5 8/7/

34 Version 0.5 8/7/

35 Summary Inputs System documentation Stakeholder interviews Processing Capturing business information requirements, business processes, and terminology Identifying and defining the source data sets Outputs Context Diagram Business Questions Candidate Source Systems Fact/Qualifier Matrix Source to Target Mapping Transformation Rules & Business Logic Glossary Business Process Diagrams Business Process Descriptions Version 0.5 8/7/

36 Questions? If you have questions, comments, or suggestions, please contact me David Loshin